We are going to set up a ~30 km link using RB-532A with R-52 radios and high quality directional 27db antennas. We get -70db signal strength when the weather is OK. Probably if the weather was worse it would have dropped a little bit. My question is what real through-put can I expect with this signal strength? Bandwidth test shows me 3.5Mbit both ways at TCP. Could this be right? I would expect a little bit more here.
If someone has any experience with similar cases please let me know.
I have 5km link and 68dbm/80dbm (dunno why they differ) but i have managed to reach with the Bandwith test 6mbits…and this is the actual speed i reached when connected all clients to the linux machine after the RB. Maybe u need some more alignement of the antennas…
One of my PTP links hovers around -68. The link is 20km 5.8ghz. I have the radios locked at 36mbps and get around 12 megs of TCP throughput. At -70 they would still be able to link at least at 24megs which should give you a sold 10 megs TCP.
My bandwith tests were done from within winbox between the 2 radios. Both CM9’s with 29dbi grids on RB532A routerboards.
that means that the actual throughput is much higher. generating the random data for the test takes all CPU resources, which drastically affects test results.
That’s one way. And like Normis said, the real throughput should be higher if it were done the right way between 2 PCs. A good way to estimate the actual throuhput is to take the wireless rate and divide it by 2. So if my radios are locked at 36mbps, the link should be able to run somewhere around 18mbps. But that’s just a rough number. Other factors can affect that number.
I just did the test again using UDP instead of TCP and the result was 17.5 mbps.